UniReal: Universal Image Generation and Editing via Learning Real-world Dynamics
This work addresses the need for a versatile solution in computer vision for image manipulation, though it appears incremental by building on video generation models.
The paper tackles the problem of diverse image generation and editing tasks by proposing UniReal, a unified framework that treats these tasks as discontinuous video generation, leveraging video data for supervision and achieving advanced capabilities in handling visual elements like shadows and reflections.
We introduce UniReal, a unified framework designed to address various image generation and editing tasks. Existing solutions often vary by tasks, yet share fundamental principles: preserving consistency between inputs and outputs while capturing visual variations. Inspired by recent video generation models that effectively balance consistency and variation across frames, we propose a unifying approach that treats image-level tasks as discontinuous video generation. Specifically, we treat varying numbers of input and output images as frames, enabling seamless support for tasks such as image generation, editing, customization, composition, etc. Although designed for image-level tasks, we leverage videos as a scalable source for universal supervision. UniReal learns world dynamics from large-scale videos, demonstrating advanced capability in handling shadows, reflections, pose variation, and object interaction, while also exhibiting emergent capability for novel applications.